减少光照影响的自适应Gamma矫正方法

在图像处理中,图像质量往往受到光照的严重影响,这给有效利用图像底层信息造成了很大的困难.为了有效弱化光照影响,本文提出了一个自适应的Gamma矫正方法.首先,我们把图像的像素值区间转换到角度值区间;然后,建立了Gamma值区间与角度值区间的非线性映射关系;最后,利用非线性Gamma矫正方法解决了Gamma值与像素值之间的对应关系.实验证明本文提出的Gamma矫正方法能有效改善图像中的光照条件,提高图像质量.     

A robust color image watermarking algorithm against rotation attacks

A robust digital watermarking algorithm is proposed based on quaternion wavelet transform (QWT) and discrete cosine transform (DCT) for copyright protection of color images. The luminance component Y of a host color image in YIQ space is decomposed by QWT, and then the coefficients of four low-frequency subbands are transformed by DCT. An original binary watermark scrambled by Arnold map and iterated sine chaotic system is embedded into the mid-frequency DCT coefficients of the subbands. In order to improve the performance of the proposed algorithm against rotation attacks, a rotation detection scheme is implemented before watermark extracting. The experimental results demonstrate that the proposed watermarking scheme shows strong robustness not only against common image processing attacks but also against arbitrary rotation attacks.     

手指静脉红外图像血管网络修复新方法

由于手指静脉位于皮下,手指中的生物组织、手指解剖结构、皮肤结构成像特性等固有原因都给手指静脉成像造成不利影响。针对手指静脉图像中普遍存在的局部血管残缺问题,首次提出一种指静脉红外图像血管网络修复方法。首先,利用多尺度Gabor滤波对手指静脉图像进行增强,减少图像整体退化性模糊;然后,对指静脉图像进行二值化并提取血管骨架网,以便对血管网络缺损位置进行判断;再将提取的血管骨架端点、二分叉点作为血管骨架网络修复的源点,根据最小路径原则实现手指静脉图像血管骨架网络修复;最后,将Gabor增强方向图作为约束条件,复原血管网络的管径信息得到修复后的手指静脉二值化图像。实验结果表明:该方法可以实现手指静脉图像局部血管网络残缺修复,得到更加完整、稳定的血管网络结构,利用修复后的图像可以进一步提高手指静脉识别精度。     

A prohibited items identification approach based on semantic segmentation

Deep learning (DL) based semantic segmentation methods can extract object information including category, location and shape. In this paper, the identification of prohibited items is regarded as a task of semantic segmentation, and proposes a universal model with automatic identification of prohibited items. This model has two improvements based on the general semantic segmentation network. Firstly, the N-type encoding structure is applied to enlarge the receptive field of the network aiming at reducing the misclassification. Secondly, consider the lack of surface texture in X-ray security images. Inspired by feature reuse in Densenet, shallow semantic information is reused to improve the segmentation accuracy. With the use of this model, when using input images of size 512×512, we could achieve 0.783 mean intersection over union (mIoU) for a seven-class object recognition problem.     

利用多特征判别的烟雾检测方法研究

基于视频图像的烟雾检测对于实现早期火灾预警具有重要意义。为了提高烟雾检测的准确率,本文提出一种利用多特征判别的烟雾检测方法。首先,采取三帧差与时域窗口累加的方法分割出烟雾候选区域,克服了传统算法无法提取稀薄烟雾前景的缺点。然后,根据烟雾的特性,提取其在单帧和多帧方面的特征,并结合烟雾本身动态变化的特点,利用马氏距离度量实现对真实烟雾区域的判别。结果表明,该方法能够在复杂环境下有效地检测烟雾,并且对稀薄烟雾起到很好的检测效果。      

陆空通话复诵语义自动化校验BiLSTM模型

为保证航空运输安全，飞行员要对管制员发送的指令进行复诵,并且管制员要对复诵指令进行进一步的确认。而由于疲劳、紧张、疏忽等原因，管制员未能及时发现飞行员复诵错误的情况也时有发生，给民航运输安全带来巨大隐患。针对这个问题，本文提出一种陆空通话复诵语义自动化校验BiLSTM模型。首先，利用两个并行的长短时记忆网络（BiLSTM）对管制员发送的指令和飞行员复诵的指令分别进行语义特征提取；然后将两个BiLSTM网络各个时刻的输出进行交互得到一个指令和复诵指令间的语义匹配矩阵；最后经过一个动态k-Max池化层后输入到多层感知器中从而得到指令与复诵指令间最终的匹配分数来判别复诵语义是否一致。实验证明，该方法在解决陆空通话复诵语义自动校验任务中是有效的，平均测试精度达到了90.53%。      

手指双模态特征图像感兴趣区域稳定定位方法研究

为了改善指静脉成像质量和稳定获取指静脉、指节纹的感兴趣区域（ROI）,本文提出了一种新的手指双模态成像系统及指静脉、指节纹ROI定位方法。首先,设计了一种含有多光谱、多亮度等级光源的手指拱起成像模型,该模型可较好改善指静脉的成像质量。然后,利用手指拱起产生的拐点实现对指节纹感兴趣区域的稳定定位,并利用双目摄像头下指静脉和指节纹成像区域之间的位置关系确定指静脉特征的感兴趣区域。实验结果表明本文所提出的方法能够较好提高指静脉血管的成像质量和指静脉、指节纹感兴趣区域的定位精度。      

Image restoration of finger-vein networks based on encoder-decoder model

Finger-vein recognition is widely applied on access control system due to the high user acceptance and convince. Improving the integrity of finger-vein is helpful for increasing the finger-vein recognition accuracy. During the process of finger-vein imaging, foreign objects may be attached on fingers, which directly affects the integrity of finger-vein images. In order to effectively extract finger-vein networks, the integrity of venous networks is still not ideal after preprocessing of finger vein images. In this paper, we propose a novel deep learning based image restoration method to improve the integrity of finger-vein networks. First, a region detecting method based on adaptive threshold is presented to locate the incomplete region. Next, an encoder-decoder model is used to restore the venous networks of the finger-vein images. Then we analyze the restoration results using several different methods. Experimental results show that the proposed method is effective to restore the venous networks of the finger-vein images.     

民航陆空通话语音识别BiLSTM网络模型

民航陆空通话对民航飞行安全十分重要,但因其通话模式有特殊的语法结构与发音方式,日常语音识别声学模型无法有效应用于民航陆空通话的语音处理问题。针对民航陆空通话的特殊语境,本文提出了基于双向长短时记忆网络(BiLSTM)的民航陆空通话语音识别方法。首先,提取民航陆空通话语音的FBANK特征作为输入,以时序链式连接(CTC)为目标函数,训练BiLSTM网络得到BiLSTM/CTC模型。然后,利用声学模型,语言模型与陆空通话词典实现民航陆空通话的语音识别,并结合数据增强与数据迁移对模型进行增强训练提高语音识别性能。实验结果表明本文提出的方法适用于民航陆空通话语音识别,并且数据增强模型可有效降低民航陆空通话语音识别的词错误率。      

X光安检图像多尺度违禁品检测

安检是保障人民生命财产安全的第一道防线,智能安检是安检行业未来发展的必然趋势。X光安检图像存在背景复杂、违禁品尺度多样以及相互遮挡现象,导致传统的目标检测算法无法获得满意的效果。本文在一阶段目标检测网络SSD框架的基础上,提出了适用于X光安检图像多尺度违禁品检测网络——非对称卷积多视野神经网络ACMNet(Asymmetrical Convolution Multi-View Neural Network)。检测网络增加了三个模块:小卷积非对称模块(Asymmetrical Tiny Convolution Module,ATM)、空洞多视野卷积模块(Dilated Convolution Multi-View Module,DCM)、多尺度特征图融合策略(Fusion strategy of multi-scale feature map,MF)。 ATM学习到的细节特征有助于小尺度违禁品的识别;DCM通过提供局部与全局之间的上下文特征信息来解决遮挡问题;MF则是通过融合高、低层特征图以提高模型在背景干扰情况下违禁品的检测精度。在仿真实验中,采用X光安检领域公开的数据集与自建的数据集,ACMNet在精确度上取得了令人满意的效果。